Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

2018.21 0fa48d9

This site may earn commission on affiliate links.
The problem is the system as is today, requires both hands on the wheel at all times to be operated safely. Not having hands on the wheel increases the time it takes for a human to take over when the AP makes a mistake.

Agreed. The truth is that none of us know the failure rate of AP in ideal conditions on a divided highway. No single person can experience enough autopilot miles to know how safe it really is. Safety can only be demonstrated by looking at millions of miles of driving data comparing hands free operation vs hand on wheel vs manual driving. Until safety is proven I’m going to keep my hands on the wheel and be attentive. The combination of attentive driver and AP is where the big safety win comes from as the technology stands today.
 
  • Like
Reactions: Helmuth
The problem is the system as is today, requires both hands on the wheel at all times to be operated safely. Not having hands on the wheel increases the time it takes for a human to take over when the AP makes a mistake.
Perhaps both hands, but at highway speeds I only use one, and I use AS all the time on highways. At speed it doesn't take much steering angle to avoid an obstacle, and for me, holding the wheel around 8 or 4 o'clock position depending on which hand, coupled with lightly gripping the wheel while occasionally gripping slightly firmer to satisfy the nag timer works well.
 
  • Like
Reactions: d21mike
Perhaps both hands, but at highway speeds I only use one, and I use AS all the time on highways. At speed it doesn't take much steering angle to avoid an obstacle, and for me, holding the wheel around 8 or 4 o'clock position depending on which hand, coupled with lightly gripping the wheel while occasionally gripping slightly firmer to satisfy the nag timer works well.
Here is the point I think most people are missing about the "hands on the wheel policy." Admittedly it took me a long time to come around but with 107,000 miles on my 2015 70D, the vast majority on AP, I have had plenty of time to come to this conclusion.

Originally I was under the impression (like most on this thread) that hands on the wheel is a poor proxy for paying attention. Just because you are holding the wheel doesn't mean you are paying attention; conversely, it is possible to be paying attention without touching the wheel.

Nag frequency is inversely proportional to autosteer path confidence. Wide-open multi-lane well-mapped clearly-lined divided highways have the least frequent nags; as confidence decreases (more traffic, curves, construction, etc.) the frequency of nags goes up. No surprise here. As confidence goes down, there is less time to react to an autosteer error (e.g., traffic nearby, in a curve).

I am suggesting that the need for hands on the wheel is more to reduce reaction time than to gauge attentiveness. As an example, I can be driving down the highway with my hand on the wheel, and become distracted (well, not ME of course...only other people). However, I will immediately notice when the wheel starts to do something I don't expect it to. This will grab my attention and I can take immediate action. The tactile feedback path is almost immediate. On the other hand, if I am sitting there, eyes glued to the road but hands on my lap, if I see the car make an incorrect move, I first have to take note of it, move my hands to the wheel, and finally take corrective action. This additional amount of time might not be sufficient in a low-confidence situation, like riding between two trucks.

So, do I like the nags? No, but now I see why holding the wheel is an effective feedback mechanism when autosteer starts to go awry.
 
10s nags has nothing to do with increased safety. F@#$ AS at that point. Tesla is saying their system is too brittle and useless to be used.
Fully agreed, it’s nothing but a BS CYA move that would cause me to drastically reduce the amount of time I spend with AS engaged. Which, if you believe Tesla’s stats about driving with AP, will actually reduce my safety, so it’s clearly not about that.

Personally I think 10 second nag intervals are too ridiculous even for Tesla, they put them in the beta knowing it would cause a stink, and now we’re all gonna feel relieved and “listened to” when the prod version hits with 15-20 second nags.
 
  • Like
Reactions: croman
Is it possible the more frequent nags are only in the odd-numbered early release builds? Perhaps they want the early release testers to be even more vigilant because of some changes that were made. It's possible that by the time the next even-numbered release is deployed to the masses, the more frequent nag won't be included.

It seems a lot of people are getting worked up about a change that only 0.2% of the cars have. Maybe wait until such a change hits the general release before worrying about it? :confused:
 
  • Like
Reactions: MP3Mike and Joe F
they need to take all that focus and energy and solve the untracked / stopped car problem, because if they can't solve that 100% of the time, none of this crap matters, because the system will never be safe.

My experience is that recognising stopped cars has improved a lot with the introduction of 10.4 and newer versions on AP 2.0 HW. I could be wrong, but weren't all accidents with hitting a stopped car or fire truck lately AP1 cars? Is it even possible to improve that nearly closed down MobilEye system and the limitation of only having 1 camera?
 
  • Like
Reactions: croman and d21mike
Fully agreed, it’s nothing but a BS CYA move that would cause me to drastically reduce the amount of time I spend with AS engaged. Which, if you believe Tesla’s stats about driving with AP, will actually reduce my safety, so it’s clearly not about that.

Personally I think 10 second nag intervals are too ridiculous even for Tesla, they put them in the beta knowing it would cause a stink, and now we’re all gonna feel relieved and “listened to” when the prod version hits with 15-20 second nags.

The increased nags is new with this version, it was not present in earlier beta's. Who knows if they will take it to production or or not. Just like the showing cars in adjacent lanes that was in 18.5 but got removed with 18.7 etc.
 
Scuttlebutt on 2018.21.1 is that its even worse than 21. Even more frequent nags (10s) and no new features or AP improvements. We're in some serious trouble. I'm avoiding installing any more updates. I wish I had stuck with 2018.12....

WADR, "scuttlebutt" reads as unsubstantiated rumor to me. Unless you know the method used by the (single?) source, or have driven with the release yourself, and never had or only had limited nags based on how you hold the wheel with earlier releases, this new 10 second rumor is meaningless.

People getting worked into a frenzy over a rumor is popcorn worthy though. :)
 
WADR, "scuttlebutt" reads as unsubstantiated rumor to me. Unless you know the method used by the (single?) source, or have driven with the release yourself, and never had or only had limited nags based on how you hold the wheel with earlier releases, this new 10 second rumor is meaningless.

People getting worked into a frenzy over a rumor is popcorn worthy though. :)
Or if you talk to someone who has it but you don't want to out your source because duh.
 
  • Like
  • Informative
Reactions: BigD0g and croman
Everything on the internet is unsubstantiated and should be taken as such.

That being said, I know my source and they have early access. Believe what you will.
No, I believe you; I don't know how your source drives their car or uses AS WRT how the steering wheel is used to provide feedback to the torque sensor.

Would you take as fact one persons report did (insert any outlandish statement of fact here) without supporting or empirical evidence? Most people would not. That's all I'm saying.

Hey, my MS 90D just did 0-60 in 1.4 seconds with the latest firmware release! I highly doubt you'd believe it.

You're right; On the 'Net, if there's no video there's no proof. Even then, if it's only one car, who's to say something isn't broken H/W wise that triggered the increased nags?

Hardly worth getting whipped into a frenzy over without proof the S/W was intentional changed to a 10 second sanity check. Even if it were, used as currently designed, varying pressure (or torque when called for) on the wheel should reset the timer such that "nags" should not happen.
 
  • Like
Reactions: croman
People aren't getting whipped in a frenzy. If you want that, go to the burger thread. This is merely lamentation about potential future AS implementations and providing fair warning to other owners to reconsider insta-updating.

Words are funny, and it's difficult to read emotion into what one writes. But, I feel that if you go back through the thread and read the increasing level of frustration in what's been written, you might feel differently if viewed in a non emotional perspective.

Do you know a hostage negotiator? If so, have them read the last few pages in this thread and give you their perspective. :)